Dipartimento di Scienze Statistiche

In the context of capture-recapture modeling for estimating the unknown size of a nite population it is often required a exible framework for dealing with a behavioural response to trapping. Many alternative settings have been proposed in the literature to account for the variation of capture probability at each occasion depending on the previous capture history. Inference is typically carried out relying on the so- called conditional likelihood approach. We highlight that such approach may, with positive probability, lead to inferential pathologies such as unbounded estimates for the nite size of the population. The occurrence of such likelihood failures is characterized within a very general class of behavioural e ect models. It is also pointed out that a fully Bayesian analysis overcomes the likelihood failure phenomenon. The overall improved performance of alternative Bayesian estimators is investigated under di erent non- informative prior distributions verifying their comparative merits with both simulated and real data.